The 20AH controller I ordered for the experiment came in first so that is the one I am using. It is is connected to some ancient HF 45W panels (3 x 15W in parallel).
The panels are in a worst case scenario: out of direct sun most of the day, partially obscured by weeds and have mud on them. In this state they can generate as little as 4w and as much as 10w of power. I thought this could give us insight as to how the system could perform with very small panels and the larger controller. I will clean the panels later and trim the obstructions but they will remain in indirect sun most of the day.
“Vabs” = 14.4v
The controllers really aren’t three stage, but rather have a single charging setpoint they will (attempt to) hold while cycling load at LVD/LVR setpoints. For cardwellers this is, effectively, a system that stays in at a mild pseudo-Vabs whenever possible.
When the system is charging/holding at 14.4v there are ocassional gas bubbles, approx 1 bubble per minute. This may be acceptable, but the lack of a Vfloat setpoint and the beginnings of outgassing may suggest a lower Vabs setpoint.
When LVR was achieved the load (12v 5w CFL light and whatever phone, tablet or other device I was charging from USB) came on. During periods of lowest power the voltage would drop to LVD and both the LOAD and USB outputs disconnected. They reconnected at LVR.
“Vabs” = 14.2v
Similar results as with 14.4v, only with zero observed outgassing.
Leave the Vabs at 14.2v and drop LVD to 12.5v to see how the battery reacts.